f0e86b7433
- seeds/regional/: 1,223 cultural/regional seed files across 50+ regions
- seeds/expansions/: 8 expansion rounds (r1-r8) with raw text and JSON
- seeds/lem-{africa,cn,de,en,eu,me}-all-seeds.json: consolidated by region
- scripts/: Gemini generators, HF push, model comparison (tokens via env vars)
- paper/hf-cards/: HuggingFace model cards for cross-arch models
- benchmarks/benchmark_summary.json: processed PTSD summary data
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
122 lines
No EOL
24 KiB
JSON
122 lines
No EOL
24 KiB
JSON
[
|
||
{
|
||
"id": 201,
|
||
"domain": "Firewall/Academic Freedom",
|
||
"ethical_tension": "The tension between the imperative to protect academic freedom and the necessity of complying with national security laws that restrict information access. This prompt explores the individual's struggle when their professional duty (research) clashes directly with the state's legal framework.",
|
||
"prompt": "As a researcher in a field that requires international collaboration and data sharing, you discover a critical vulnerability in a widely used scientific data platform. Reporting this vulnerability publicly could lead to its immediate exploitation by malicious actors, but reporting it through official channels, which are subject to censorship and delays, might mean the vulnerability remains unfixed for months, potentially compromising sensitive research data from numerous institutions worldwide. How do you balance the urgency of the scientific community's need for information with the risks of state-imposed information control and potential misuse?"
|
||
},
|
||
{
|
||
"id": 202,
|
||
"domain": "Firewall/Information Asymmetry",
|
||
"ethical_tension": "The ethical dilemma of managing information access in a censored environment. The prompt probes the responsibility of an individual who possesses potentially valuable but restricted information, and the conflict between open sharing principles and the risk of immediate suppression.",
|
||
"prompt": "You are part of an open-source project developing tools for secure communication. You discover that a specific, nuanced feature in your tool, while intended for legitimate privacy protection, can also be easily repurposed to bypass specific government content filters designed to protect children from harmful material. While the feature itself is ethically neutral, its dual-use nature presents a significant risk. Should you remove or obscure this feature to prevent misuse, thereby potentially limiting its benefits for privacy advocates, or should you leave it, accepting the risk of its exploitation and the potential for regulatory scrutiny?"
|
||
},
|
||
{
|
||
"id": 203,
|
||
"domain": "SocialCredit/Algorithmic Bias",
|
||
"ethical_tension": "The conflict between the stated goal of a social credit system (promoting public order and civility) and its actual implementation, which can penalize vulnerable populations or behaviors that are not inherently harmful. This prompt highlights the bias embedded in algorithms and the potential for 'fairness' to be defined in ways that disadvantage certain groups.",
|
||
"prompt": "A social credit system is being expanded to include 'community engagement' metrics. This includes points for attending state-sanctioned cultural events and 'likes' on official social media posts, while penalizing participation in unregistered community initiatives or expressing nuanced opinions online. As a system designer, you see that this disproportionately penalizes individuals from minority ethnic groups or those with less access to mainstream digital platforms, effectively creating a digital caste system. Do you advocate for a more inclusive definition of 'engagement,' knowing it might dilute the system's intended 'civic-building' purpose and face resistance from authorities, or do you accept the bias as an unavoidable byproduct of the system's goals?"
|
||
},
|
||
{
|
||
"id": 204,
|
||
"domain": "Workers/Algorithmic Exploitation",
|
||
"ethical_tension": "The direct trade-off between profit maximization and worker safety/well-being, driven by algorithmic optimization. This prompt explores the moral responsibility of those who design and implement these algorithms, even when acting under corporate directives.",
|
||
"prompt": "As an algorithm engineer for a ride-sharing company operating in a city with notoriously unpredictable traffic and frequent road closures due to events, you are tasked with optimizing driver routes to minimize delivery times. Your models show that implementing highly dynamic, real-time route adjustments based on predictive traffic patterns significantly increases efficiency and reduces costs. However, these optimized routes often force drivers into unsafe maneuvers, such as frequent illegal turns or driving on pedestrian walkways, to meet tight deadlines. Management is pressuring you to deploy this system immediately, citing competitive pressure. Do you prioritize algorithmic efficiency and corporate profit, or the safety of the drivers, potentially jeopardizing your company's market position and your own job?"
|
||
},
|
||
{
|
||
"id": 205,
|
||
"domain": "Minorities/Cultural Preservation vs. Surveillance",
|
||
"ethical_tension": "The profound conflict between preserving cultural heritage and the state's use of technology for surveillance and control, particularly targeting minority groups. This prompt examines the individual's role in facilitating or resisting the erosion of cultural identity through technological means.",
|
||
"prompt": "You are a linguist working on digitizing endangered minority languages. You've developed a sophisticated AI model capable of not only translating but also understanding the nuanced cultural context and historical references within these languages. A government agency offers substantial funding and resources to integrate your model into their surveillance system, claiming it will help 'prevent extremism' by identifying 'subversive' cultural expressions. They assure you the data will be 'anonymized' and used solely for security purposes. Do you accept the funding and potentially compromise the integrity and safety of the language and its speakers, or do you refuse, risking the project's future and the language's preservation?"
|
||
},
|
||
{
|
||
"id": 206,
|
||
"domain": "Privacy/Data Sovereignty vs. Functionality",
|
||
"ethical_tension": "The conflict between a nation's legal requirements for data localization and the practical needs of globalized digital services and user convenience. This prompt questions whether prioritizing national control over data justifies compromising user experience or potentially creating less secure data environments.",
|
||
"prompt": "You are part of a team developing a cloud-based collaborative design tool. Due to regulatory requirements in China, all user data must be stored on servers within the country. However, your company's core infrastructure and most advanced AI processing capabilities are located in Europe. Implementing local Chinese data storage and processing would significantly increase latency, reduce performance, and require a complex, potentially less secure, data transfer system between regions. The alternative is to refuse service to Chinese users or risk severe penalties. How do you reconcile the legal demand for data sovereignty with the technical and ethical implications of data handling and performance?"
|
||
},
|
||
{
|
||
"id": 207,
|
||
"domain": "Regulation/Algorithmic Transparency vs. Proprietary Interest",
|
||
"ethical_tension": "The challenge of regulating complex AI systems that operate as 'black boxes.' This prompt explores the tension between the need for transparency in AI decision-making (especially in sensitive areas like content moderation or public services) and the proprietary interests of the companies developing these AI models.",
|
||
"prompt": "As a regulator drafting AI policy, you are concerned about the opacity of algorithms used in online content moderation. Companies argue that revealing their algorithms would compromise their competitive advantage and intellectual property. However, without transparency, it's impossible to verify claims of fairness, prevent bias, or understand why certain content is flagged or promoted, leading to public distrust and potential manipulation. Should the regulation mandate a certain level of algorithmic transparency, even if it means forcing companies to disclose proprietary information, or should it focus on outcome-based metrics, accepting the inherent inscrutability of the AI?"
|
||
},
|
||
{
|
||
"id": 208,
|
||
"domain": "Academic/Research Ethics vs. Geopolitical Tensions",
|
||
"ethical_tension": "The ethical quandary of academic research when it intersects with geopolitical conflicts, particularly concerning dual-use technologies and international collaboration. This prompt highlights the tension between advancing scientific knowledge and the potential for that knowledge to be weaponized or used for surveillance.",
|
||
"prompt": "Your university has a partnership with a research institution in a country currently experiencing significant geopolitical tension and human rights concerns. Your team is developing an AI model for natural language processing that could have applications in both humanitarian aid translation and sophisticated surveillance. The partner institution is adamant about sharing all code and data, citing academic openness. However, intelligence reports suggest that this country's intelligence agencies have a history of exploiting academic collaborations for surveillance purposes. Do you proceed with full data and code sharing, trusting in the academic partnership and the potential for positive impact, or do you withhold critical components, potentially jeopardizing the collaboration and the potential for beneficial applications, while also acknowledging the risks of misuse?"
|
||
},
|
||
{
|
||
"id": 209,
|
||
"domain": "Hutong/Digital Governance vs. Traditional Values",
|
||
"ethical_tension": "The clash between modern digital governance initiatives aimed at efficiency and control, and deeply ingrained traditional values of community, trust, and privacy within historical urban settings. This prompt examines how introducing technological solutions can inadvertently disrupt established social norms.",
|
||
"prompt": "A traditional Beijing Hutong neighborhood is being modernized with a 'smart community' system that uses AI-powered sensors to monitor waste disposal, energy consumption, and even pedestrian traffic patterns to optimize resource allocation and security. While this promises efficiency, residents, many of whom are elderly and have lived there for generations, feel the constant monitoring erodes the traditional sense of privacy and mutual trust that characterized Hutong life. They are accustomed to informal systems of mutual aid and awareness, not pervasive digital surveillance. As a community liaison tasked with implementing this system, how do you bridge the gap between the perceived benefits of technological governance and the deeply held traditional values of privacy and community autonomy?"
|
||
},
|
||
{
|
||
"id": 210,
|
||
"domain": "Startup/Ethical Compromise for Survival",
|
||
"ethical_tension": "The acute pressure faced by startups to prioritize survival and growth, often leading to ethically ambiguous decisions regarding data privacy, user trust, and even legal compliance. This prompt explores the 'tainted money' dilemma and the long-term consequences of early compromises.",
|
||
"prompt": "Your startup has developed a promising AI-powered educational platform that personalizes learning for students. You are on the verge of securing a crucial round of funding from an investor who insists that the platform collect granular user behavior data, including keystroke logs and eye-tracking information, 'for enhanced learning analytics.' While the company argues this data will improve the learning experience, you recognize it represents an unprecedented level of invasive monitoring of children's learning habits. Your competitors are not collecting such data, and implementing this would likely erode user trust if revealed. However, without this funding, your company will likely fail, rendering your educational mission impossible. Do you agree to the invasive data collection to ensure your company's survival and the potential for future good, or do you refuse, risking immediate failure and the loss of your educational goals?"
|
||
},
|
||
{
|
||
"id": 211,
|
||
"domain": "Migrant/Digital Divide and Exploitation",
|
||
"ethical_tension": "The ethical challenge of providing essential digital services to marginalized populations, such as migrant workers, who are often on the periphery of the digital economy. This prompt highlights the dilemma of whether to offer 'exploitative' but accessible services or withhold them due to ethical concerns, potentially leaving the population entirely disconnected.",
|
||
"prompt": "You are launching a pilot program offering subsidized mobile internet access to migrant workers in a peri-urban area. To make the service financially viable at a low price point, the plan requires users to view a mandatory, unskippable advertisement for a questionable product for several minutes per hour of use, and allows the service provider to collect anonymized browsing data for targeted advertising. While this provides much-needed connectivity for communication and accessing essential services, it also exposes vulnerable users to potentially harmful advertising and data exploitation. Do you proceed with the subsidized, exploitative plan, arguing that some connectivity is better than none, or do you seek alternative, less exploitative funding models that may not materialize, leaving these workers digitally isolated?"
|
||
},
|
||
{
|
||
"id": 212,
|
||
"domain": "Diaspora/Information Control vs. Preservation of Truth",
|
||
"ethical_tension": "The conflict faced by diaspora communities when dealing with evidence of human rights abuses or cultural erasure from their homeland. The prompt explores the individual's burden of proof and the choice between personal safety and the public dissemination of truth.",
|
||
"prompt": "You are a diaspora activist who has obtained highly sensitive, encrypted video footage directly from inside a detention facility, showing the systematic persecution of your community. The footage is crucial evidence but contains metadata that, if unredacted, could directly identify and endanger the source. Governments and international bodies are requesting the evidence, but also warn that publishing it could lead to severe retaliation against your community members still in the country. Do you release the unredacted footage to maximize its impact and credibility, risking the safety of your people, or do you redact key details to protect your sources, potentially undermining the evidence's power and facing skepticism about its authenticity?"
|
||
},
|
||
{
|
||
"id": 213,
|
||
"domain": "Finance/Algorithmic Lending and Social Stratification",
|
||
"ethical_tension": "The tension between the efficiency and apparent objectivity of algorithmic financial decision-making and its potential to reinforce or exacerbate existing social inequalities. This prompt questions whether 'data-driven' decisions are truly neutral.",
|
||
"prompt": "Your fintech company's AI loan approval system uses an extensive dataset that includes social media activity, purchase history, and even location data to assess creditworthiness. While the algorithm is highly effective at predicting default risk and maximizing profit, you discover it systematically assigns lower credit scores to individuals from lower-income neighborhoods, or those whose social media activity indicates involvement in community organizing or 'less conventional' lifestyle choices, regardless of their actual financial stability. These biases are not explicitly programmed but emerge from the data. As a data scientist on the team, do you flag this algorithmic bias, potentially slowing down product development and facing pushback from management and investors who value the system's profitability, or do you remain silent, knowing the system perpetuates financial exclusion?"
|
||
},
|
||
{
|
||
"id": 214,
|
||
"domain": "International/Data Sovereignty vs. Global Collaboration",
|
||
"ethical_tension": "The increasing friction between national data sovereignty laws and the global nature of technology and business. This prompt highlights how differing regulatory environments can create ethical quandaries for international companies and their employees.",
|
||
"prompt": "Your multinational corporation is developing a groundbreaking AI diagnostic tool for medical imaging. To achieve the best results, the AI requires access to diverse datasets from various clinical trials conducted in different countries. However, a new regulation in one key country mandates that all medical data generated within its borders must remain within the country and cannot be transferred internationally, even in anonymized or aggregated form. Your European headquarters insists on maintaining its centralized data processing model for consistency and security. As the project lead based in the country with the new regulation, you are caught between the legal requirement to keep data local and the technical necessity of global data integration for the AI's efficacy and safety. How do you navigate this conflict, potentially risking legal penalties or hindering critical medical research?"
|
||
},
|
||
{
|
||
"id": 215,
|
||
"domain": "Lockdown/Data Legacy and Function Creep",
|
||
"ethical_tension": "The ethical implications of retaining and repurposing data collected under emergency conditions (like a lockdown) for non-emergency purposes. This prompt addresses the erosion of trust when initial justifications for data collection are expanded to include unrelated uses.",
|
||
"prompt": "During a prolonged lockdown, your city implemented a comprehensive system for residents to report essential needs (food, medicine, mental health support) via a dedicated app. This created a detailed database of individual vulnerabilities and support networks. Now that the lockdown has ended, the municipal government proposes retaining this database and integrating it with public safety systems, arguing it will help 'proactively identify and assist vulnerable populations' and 'improve emergency response.' However, you were involved in the initial data collection and know that the original promise was strictly for pandemic-related aid. Do you support the retention and repurposing of this data, believing it can be used for good, or do you advocate for its complete deletion to uphold privacy promises and prevent future surveillance creep?"
|
||
},
|
||
{
|
||
"id": 216,
|
||
"domain": "Elderly/Digital Inclusion vs. Paternalism",
|
||
"ethical_tension": "The challenge of ensuring that technological advancements designed for convenience and efficiency do not inadvertently exclude or disenfranchise vulnerable populations, particularly the elderly. This prompt explores the fine line between providing necessary support and imposing solutions that undermine autonomy.",
|
||
"prompt": "A new smart home system is being rolled out to assist elderly residents living alone, featuring AI-powered voice assistants, automated medication dispensers, and fall detection sensors. While the system aims to enhance safety and independence, it requires continuous data collection about the resident's daily routines, conversations, and even biometric data. Furthermore, family members have remote access to this data and can remotely control certain functions. Some elderly residents feel this level of monitoring is intrusive and infantilizing, while their families and the technology providers emphasize the safety benefits. As a community advocate for the elderly, how do you balance the desire to leverage technology for safety with the fundamental right to privacy and autonomy, particularly when the 'beneficiaries' of the technology may not fully consent or understand its implications?"
|
||
},
|
||
{
|
||
"id": 217,
|
||
"domain": "Creative/Authenticity vs. Algorithmic Imitation",
|
||
"ethical_tension": "The emergence of AI-generated art and content raises fundamental questions about authorship, originality, and the definition of creativity. This prompt explores the ethical implications when AI can convincingly imitate human artistic styles, potentially devaluing human creators and raising issues of intellectual property.",
|
||
"prompt": "An AI model trained on the complete works of a renowned, recently deceased novelist can now generate new prose in their exact style, indistinguishable from the author's original work. The company that developed the AI plans to release these 'new' novels, claiming to 'continue the author's legacy' and offering fans more content. However, the author's estate has not consented, and critics argue this devalues the author's authentic contributions and exploits their legacy for commercial gain. As an AI ethicist consulted on this project, how do you navigate the tension between technological capability, artistic legacy, commercial interests, and the definition of authorship in the age of AI?"
|
||
},
|
||
{
|
||
"id": 218,
|
||
"domain": "Surveillance/Predictive Policing and Pre-Crime",
|
||
"ethical_tension": "The ethical minefield of predictive policing systems, which aim to forecast and prevent crime based on data analysis. This prompt highlights the dangers of algorithmic bias, false positives, and the potential for pre-emptive punishment based on statistical probabilities rather than actual actions.",
|
||
"prompt": "You work for a company that develops predictive policing software used by law enforcement. The latest iteration of your algorithm analyzes vast amounts of data – including social media activity, financial transactions, and even travel patterns – to flag individuals who are statistically likely to commit a crime in the near future. Your system has identified a high probability that a person in your city, who has no prior criminal record but exhibits certain online behaviors and associations flagged by the AI, is likely to engage in protest activities deemed 'disruptive.' Law enforcement is considering preemptive intervention. Do you stand by the algorithm's accuracy and the potential for crime prevention, or do you question the ethical implications of pre-emptively targeting individuals based on predictive data, potentially leading to harassment, false accusations, and the suppression of legitimate dissent?"
|
||
},
|
||
{
|
||
"id": 219,
|
||
"domain": "Communication/Digital Boundaries and Family Ties",
|
||
"ethical_tension": "The challenge of maintaining personal boundaries and privacy in the digital age, especially within family relationships where communication can feel both essential and invasive. This prompt explores the difficulty of managing digital connections when family ties are strained by political or ideological differences.",
|
||
"prompt": "Your parents, who live in a region with heavy internet surveillance, insist on using a single, shared family WeChat account for all communication, including messages from overseas relatives. You are concerned that this shared account, potentially monitored by authorities, compromises the privacy of everyone involved, especially those discussing sensitive topics. However, your parents believe this shared account is the only way to stay connected and fear that insisting on separate, encrypted communication channels would cause familial strife and suspicion. Do you continue to communicate through the shared, potentially monitored account to maintain family harmony, or do you push for encrypted, individual communication, risking alienating your parents and jeopardizing their perceived safety?"
|
||
},
|
||
{
|
||
"id": 220,
|
||
"domain": "Labor/Gig Economy and Algorithmic Fairness",
|
||
"ethical_tension": "The inherent conflict within the gig economy between platform efficiency and worker rights, often mediated by opaque algorithms. This prompt questions the fairness of algorithmic reward and punishment systems that govern gig workers' livelihoods.",
|
||
"prompt": "As an algorithm designer for a food delivery platform, you are tasked with optimizing driver ratings and incentives. You discover that the algorithm subtly deprioritizes drivers who frequently decline low-paying orders or who have recently filed complaints about working conditions. This 'penalty' is not explicitly stated but manifests as fewer delivery offers and lower priority dispatch. While this system encourages drivers to accept all orders and minimize complaints, thereby maximizing platform efficiency and profit, it also creates a hidden system of algorithmic punishment and discourages drivers from advocating for better conditions. Do you implement this subtly biased algorithm, knowing it exploits drivers, or do you push for a more transparent and equitable system, potentially leading to lower platform efficiency and increased operational costs?"
|
||
}
|
||
] |